Support trying to parse `<citation>` as `<element-citation>`

substance / texture

A visual editor for research.

MIT License

1k stars 83 forks source link

Support trying to parse `<citation>` as `<element-citation>` #944

Open axfelix opened 5 years ago

axfelix commented 5 years ago

A significant part of the Pubmed Central JATS corpus has <citation> elements that would be complete if they were parsed as <element-citation> -- that is, a simple find-and-replace is sufficient to make the document validate against Dar. It would be nice if Texture supported this.

Melissa37 commented 5 years ago

Hi @axfelix would you be able to provide an example of what you mean? I don't understand, sorry!

axfelix commented 5 years ago

Sure. Right now, Texture only explicitly supports documents with citations encoded as <element-citation> rather than <citation>, I assume because the latter can be messier. However, if you look at the PubMed Central JATS corpus (one of the largest ones around, surely), many documents have citations encoded as regular <citation> but whose child elements are complete if the tag were simply find-and-replaced to <element-citation>, so I think we should be able to support these on import.

Melissa37 commented 5 years ago

Ah, thanks, got you now. We discussed this a fair bit at the last meet up. This is a requirement for Erudit. For eLife we have structured references and are very strict, but for people using this as a post production tool they cannot control the input and so need flexibility for string references.

fabiobatalha commented 5 years ago

Hi @axfelix

Do you mean <mixed-citation> instead of <citation>?

The <citation> element is not available in JATS.

Related discussion #412

obuchtala commented 5 years ago

Thanks. As @Melissa37 said, we have discussed this in Sao Paolo.

To summarise, not being able to have 'unstructured' citations, is preventing Erudit from adopting Texture in their process. To address this, we will introduce a mode that allows (and requires!) editing references and citations manually. And we hope that this is needed only as an interim solution.